2^n元数神经网络模型：多值图象的识别应用

The 2￣n-element Number Neural Networks Models:Application of the Multistate Patterns Recognition

对２￣ｎ元数神经网络模型的计算机数值模拟结果表明，其存贮容量和纠惜能力随ｎ的增大而降低，讨论了四态复数模型或十六态四元数模型的四态或十六态多值图象的识别应用实例，并对其光学实现进行了初步的探讨．

In this paper,the computer numerical simulation results show that the stor-age capacity and the error-correct ability of the 2￣n-element number neural network models de-crease with the increase of n.Some examples of the 4 or 16 states patterns recognition with n=1 or n=2 are discussed.Its optical implementation is simply discussed.